The quest for higher speed of operations has been the trend for last two decades and there was not much research efforts to find the right device solutions for the low speed digitals devices with really low sleep power consumption.
I get the impression that Ambiq is Cortex-M0+ development while ARM is doing next-gen core research.
NEAR versus SUB
But note well ARM's processor core is being optimized for energy harvest type use in a NEAR-threshold implementation. This makes sense if ARM wants TSMC to provide support with a process flow and be able to license the core widely.
On the other hand Ambiq is aggressively going after SUB threshold because they are prepared to back their design skills and take on the risk of working outside TSMC's design flow envelope.
That is going to depend as much on the lawyers as the engineers. If they have good patent protection then they should be OK, but the fact that Arm Holdings funded them probably means that they are keeping an eye on the technology. I could see them getting bought out if Arm decides to integrate it as core technology.
Did they say if there is a penalty on switching speed or accuracy with this technology? It seems like there would be some kind of price to be paid for the increased power efficiency.